Montrose et al.

Allergy Asthma Clin Immunol (2017) 13:12 Allergy, Asthma & Clinical Immunology
DOI 10.1186/s13223-017-0187-8

RESEARCH Open Access

Dietary intake is associated

with respiratory health outcomes and DNA
methylation in children with asthma
L. Montrose1, T. J. Ward2, E. O. Semmens2, Y. H. Cho2, B. Brown3 and C. W. Noonan2*

Abstract 
Background:  Asthma is an increasingly common chronic disease among children, and data point toward a complex
mechanism involving genetic, environmental and epigenetic factors. Epigenetic modifications such as DNA hypo- or
hyper-methylation have been shown to occur in response to environmental exposures including dietary nutrients.
Methods:  Within the context of the asthma randomized trial of indoor wood smoke (ARTIS) study, we investigated
relationships between diet, asthma health measures, and DNA methylation. Asthma health measures included a
quality of life instrument, diurnal peak flow variability (dPFV) and forced expiratory volume in the first second (FEV1).
Dietary intake was assessed with a food frequency questionnaire. Methylation levels of LINE-1 repetitive element and
two promoter CpG sites for interferon gamma (IFNγ, -186 and -54) from buccal cell DNA were measured using pyrose-
quencing assays.
Results:  Data were collected on 32 children with asthma living in western Montana who were recruited to the ARTIS
study. Selenium and several methyl donor dietary nutrients were positively associated with the asthma quality of life
measure. Intake of methyl donating nutrients including folate was positively associated LINE-1 methylation and nega-
tively associated with IFNγ CpG-186. Higher levels of LINE-1 methylation were associated with greater dPFV.
Conclusion:  We identified several nutrients that were associated with improved quality of life measures among
children with asthma. The IFNγ promoter CpG site -186 but not -54 was associated with the intake of selected dietary
nutrients. However, in this small population of children with asthma, the IFNγ promoter CpG sites were not associ-
ated with respiratory health measures so it remains unclear through which epigenetic mechanism these nutrients
are impacting the quality of life measure. These findings add to the evidence that dietary nutrients, particularly foods
containing methyl donors, may be important for epigenetic regulation as it pertains to the control of asthma.
Trial registration ClincialTrials.gov NCT00807183. Registered 10 December 2008
Keywords:  Asthma, Methylation, Spirometry, Diet, Nutrition, Children, Epigenetics, Quality of life

Background countries [2]. Epidemiological studies suggest that die-

Asthma is an environmentally triggered disease that
affects nearly 26 million people in the United States [1].
Dietary intake represents a modifiable environmental
exposure that could partially explain the current burden
of chronic disease, including asthma, in industrialized
*Correspondence: curtis.noonan@mso.umt.edu
2
Center for Environmental Health Sciences, University of Montana, 32
Campus Drive‑159 Skaggs, Missoula, MT 59812, USA
Full list of author information is available at the end of the article
tary patterns are linked to the risk of developing asthma,
however, the evidence from longitudinal birth cohorts
has not clearly defined the importance of specific nutri-
ents or fully elucidated the mechanistic pathways linking
diet to chronic respiratory disease. Further, there have
been few studies aimed at determining if nutrient intake
contributes to asthma control in children. One potential
mechanism whereby dietary intake affects respiratory
health in children is through epigenetic modulation of
immunoregulatory cytokines.

© The Author(s) 2017. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License
(http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium,
provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license,
and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/
publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Montrose et al. Allergy Asthma Clin Immunol (2017) 13:12
Significant observational data suggests that dietary
status and intake of particular nutrients can affect res-
piratory health outcomes. Several recent studies have
suggested that some dietary nutrients may be protective
for respiratory health [3–12]. A recent review concluded
that dietary intake in utero and throughout the lifecourse
can influence respiratory health status, however defini-
tively assessing causal relationships in human studies is a
major challenge [13]. A cross-sectional study by Berthon
et al. showed that among asthmatics, a high fat diet was
associated with increased airway eosinophilic inflamma-
tion, and low fiber intake was associated with poor lung
function [14]. Supplementation of dietary folic acid has
been successful in the prevention of neural tube defects
in the United States. However, longitudinal cohort stud-
ies have produced mixed results regarding maternal
folic acid supplementation and asthma development [15,
16]. Antioxidants like selenium may play a role in res-
piratory health through systemic reduction of oxidative
stress [17]. In a mouse model of allergic airway disease, a
combinatory therapeutic that included selenium attenu-
ated the physiologic airway damage that is typical of this
model [18].
The rapidly evolving field of epigenetics has emerged
as an appealing potential mechanistic bridge that could
link environmental exposures to the development of
asthma or the exacerbation of asthma-related symptoms
[19]. The exact toxicoepigenetic mechanisms are far from
elucidated, but landmark studies using the agouti mouse
model have provided solid evidence that environmental
exposures can affect phenotype through alterations in
DNA methylation patterns [20]. Understanding how and
when these mechanisms can impact asthma pathogenesis
is paramount. In a mouse model of allergic airway dis-
ease, in utero dietary intake of methyl donating nutrients
was associated with an enhanced disease phenotype as
well as aberrant hypermethylation of runt-related tran-
scription factor 3 (Runx3), a gene known to suppress
allergic airway disease [21]. Although the perinatal expo-
sure window may be especially important, data also sug-
gest that environmental exposures could impact health
via epigenetic mechanisms throughout the lifecourse.
In humans, the production of regulatory T cells (Tregs),
which are known to suppress immune responses, is con-
trolled by transcription factor forkhead box p3 (FOXP3)
[22]. Nadeau et  al. demonstrated that patients with
asthma in a polluted environment had a hypermethylated
FOXP3 locus profile which was associated with impaired
Treg function relative to patients with asthma in a less
polluted area [23].
The relationship between dietary intake and epige-
netic modifications is complex and compounded by sen-
sitivity to timing of exposure (e.g. prenatal, postnatal,
Page 2 of 12
adolescent, or adult). Nevertheless, human and mouse
data indicate several dietary nutrients play a role in epi-
genetic mechanisms [24], thus it is possible that nutri-
ent intake is related to asthma pathogenesis through the
epigenetic regulation of key genes. Asthma is pheno-
typically characterized by a shift toward type 2 T helper
(Th2) polarization and consequently type 1 T helper
(Th1) cell cytokines such as interferon gamma (IFNγ)
play a critical role as counter regulators in the allergic
asthma pathway [25, 26]. For example, in a follow-up
study of adults recruited as children with a history of
wheeze, those with persistent asthma were compared to
those with resolved asthma to characterize the Th1/Th2
response following exposure to house dust mite allergen
[27]. Smart et al. found that those with persistent asthma
had much weaker Th1 responses and concluded that a
measured decrease in IFNγ production in this group
could be a major factor underpinning the presence of
severe and chronic asthma symptoms. Meng et al. inves-
tigated the effect of diet on IFNγ production in humans
and showed that cells extracted and purified from non-
asthmatic adults produced differential amounts of IFNγ
[28]. Interestingly, Meng found that the amounts of IFNγ
were associated with intake of specific dietary variables
and predicted upper respiratory tract infection incidence.
Finally, a series of studies using either a ragweed or dust
mite-sensitized mouse model of asthma showed that pre-
treatment with a DNA adjuvant known to result in Th1
biased immune status with marked overproduction of
IFNγ resulted in an ameliorated lung inflammatory phe-
notype [29, 30]. Thus IFNγ is a relevant candidate gene
that plausibly exists in the mechanistic pathway linking
dietary intake to respiratory health via epigenetic regula-
tion of the Th1/Th2 cytokine balance.
Poor asthma control is associated with school absences,
higher health care costs and worse long-term health out-
comes. An understanding of the association between
a child’s recent dietary history and respiratory health
measures could lead to important intervention strategies
to improve outcomes among children with asthma. In
this study we aimed to evaluate the relationship between
a priori selected nutrients and asthma health. Although
the link between current dietary status and asthma
health is not clear, evidence suggests a potential role for
an epigenetic mechanism. In addition to a measure of
global gene methylation, IFNγ was chosen as a candidate
gene because of its well-established role in the Th1/Th2
balance.
Methods
Study overview
Participants were recruited from the asthma randomized
trial of indoor wood smoke (ARTIS) study. The rationale
Montrose et al. Allergy Asthma Clin Immunol (2017) 13:12
and methods for the ARTIS study have been described
previously [31–33]. The ARTIS study included 114 chil-
dren with asthma (ages 6–17) from 97 homes in Mon-
tana, Idaho, and Alaska. This parent study was designed
to test an indoor air quality intervention, and homes were
assigned to either a placebo arm or an air filter interven-
tion. Two in-home data collection visits occurred in each
of two consecutive winter periods with the intervention
occurring between these winter periods. The subco-
hort recruited to participate in this diet and epigenetics
study included 32 participants living in western Montana
who had been recruited in the final 2 years of the 5-year
ARTIS study. Additional file 1: Figure S1 indicates when
spirometry measurements, buccal cells, and food fre-
quency questionnaires (FFQ) were administered. For the
purpose of the currently described study, only data that
was collected in conjunction with a FFQ was considered.
In Additional file 1: Figure S1, this would be visits B and
D. Health outcomes measures included a quality of life
instrument and self-monitoring of spirometry measures
using a peak flow meter. Buccal cell samples were col-
lected for evaluating epigenetic markers. Anthropomet-
ric measures determined by trained staff using a digital
scale and stadiometer along with the participant’s gen-
der and date of birth were used to calculate body mass
index (BMI) percentile using the U.S. Centers for Disease
Control and Prevention (CDC) calculator [34]. The study
was approved by the University of Montana Institutional
Review Board. In addition to the informed consent pro-
cedures for the parent study, children were separately
assented to participate in this diet and epigenetics study
and parents signed a parental permission and informed
consent form.
Dietary nutrient collection
Dietary data was collected using the 2004 Block Kids
FFQ (NutritionQuest, Berkeley, CA, USA) to charac-
terize dietary intake among participants. This instru-
ment has been validated in children, ages 6–17  years
old [35–38]. The questionnaire includes 77 food items.
In addition to intake of standard nutrients, this instru-
ment was used to estimate intake of micronutrients that
participate in the one-carbon metabolism pathway (i.e.
betaine, choline, folate, etc.). These nutrients are impor-
tant for the generation of methyl groups and are there-
fore potentially relevant to DNA methylation markers.
The questionnaire was administered to each participant
by trained staff using serving size visual aides, and par-
ents were asked to assist their child with portion size rec-
ognition and remembering foods they ate during the last
week. Questionnaires were processed by NutritionQuest,
and the resulting data were analyzed at the University of
Montana.
Page 3 of 12
Health outcome measures collected in parent study
The pediatric asthma quality of life questionnaire
(PAQLQ) is a 23-item asthma-specific battery which
provides domain scores for symptoms (10 items), activ-
ity limitation (5 items), and emotional function (8 items)
[39]. The total PAQLQ score and each domain score are
calculated as mean scores ranging from one to seven with
seven as the optimal score. The PAQLQ has been vali-
dated as an evaluative tool to measure within participant
changes over time due to treatment, and changes in this
scale of 0.5 or more points are clinically significant [39].
Using the PiKo-1  m (Ferraris Respiratory, Ayer, MA,
USA) participants performed a test twice daily, in the
morning and in the evening, for a period of 2 weeks.
These 2-week periods were initiated at the beginning of
each air sampling event. For each test, the child’s parent
records the observation as it appears on the meter and
these observations are later checked for accuracy against
the digital log of the instrument. The instrument records
the best result for both peak expiratory flow (PEF) and
forced expiratory volume in one second (FEV1). Out-
comes from these measures include average morning
PEF and FEV1, average evening PEF and FEV1, and diur-
nal PEF variability (dPFV).
Cell collection, DNA extraction, and pyrosequencing
Buccal cells were collected from the participant’s cheek
by trained staff using a cytology brush and stored in Cell
Lysis Solution (Qiagen, Valencia, CA, USA) at room tem-
perature until all samples were collected. In compliance
with this protocol, all samples were processed within
24  months from the day of collection. DNA from the
buccal cells was extracted using Gentra Puregene Buccal
Cell DNA Kit (Qiagen, Valencia, CA, USA) according to
the manufacturer’s instructions. The quantity of the puri-
fied DNA was measured using a Nanodrop spectropho-
tometer (Thermo Scientific, Wilmington, DE) and then
stored at −20  °C. DNA bisulfite treatment was carried
out using the EZ DNA Methylation-Direct Kit (Zymo
Research, Irvine, CA) according to the manufacturer’s
instruction, and stored at −20 °C. Pyrosequencing assay
was used to measure methylation levels of LINE-1 repeti-
tive element and the promoter region of INFγ. Briefly,
50  µg of bisulfite-modified DNA was PCR amplified by
polymerase chain reaction (PCR) using specific primers
(Additional file  1: Table S1) and the PyroMark PCR kit
(Qiagen, Valencia, CA, USA). After annealing, pyrose-
quencing was conducted using a Pyromark Q96 MD
(Qiagen, Valencia, CA, USA). Samples were run in dupli-
cate and only samples with a coefficient of variation less
than 5% were used in the final analysis. Epitect (Qiagen,
Valencia, CA, USA) bisulfite treated controls, which
include a methylated and unmethylated human genome
Montrose et al. Allergy Asthma Clin Immunol (2017) 13:12 Page 4 of 12

sample, along with a no template control were used on buccal cells, PAQLQ and spirometry and therefore, only
each plate. these ‘complete’ visits were considered in the analysis.
Approximately 63% of subcohort, or 20 participants, had
Statistical analysis both a year one and year two ‘complete’ visit, while 12
All analyses were conducted using SAS v9.4 (Cary, NC, participants only had one ‘complete’ visit, which occurred
USA). To evaluate if cross sectional measurements of in either year one or year two, for a total of 52 observa-
IFN CpG sites are correlated with each other and/or tions. Reasons for these 12 participants having one rather
correlated with LINE1 global methylation we estimated than 2  years of data included missing data, participant
Pearson correlation coefficients using the first available not available during scheduled visit, or the participant
observation for each participant (n  =  32). A subset of chose to only participate in one year of the study. Moreo-
17 macro- and micronutrients from the total 73 nutrient ver, in the final GEE models, which were adjusted for sev-
variables generated by the FFQ were chosen after a litera- eral covariates, one participant (two observations) was
ture review of diet as it relates to asthma. Relationships excluded because income and education data was miss-
between a priori selected dietary nutrients and both epi- ing, therefore the results from these models include 50
genetic markers and asthma outcomes were considered observations from 31 participants. Ages ranged from 8 to
in separate models using all available and complete data, 17  years and 47% were male (Table  1). The study popu-
which included multiple visits for some participants. lation was 94% non-Hispanic white. The mean (sd) BMI
Their associations with continuous epigenetic markers percentile was 70.6 (20.1) and 34% (n =  11) were above
(i.e. global and gene-specific methylation) and asthma the 85th percentile, which is considered overweight
measures were evaluated using generalized estimating according to the CDC [34]. Baseline asthma-related res-
equations (GEE), which account for correlations between piratory health values can be found in Table  1. Mean
repeated measures on the same participant. Tertiles of values for both dPFV and FEV1 were at the approximate
dietary nutrients were included in analyses as three-level threshold used to designate poor asthma control [41].
indicator variables to investigate potentially non-linear Mean (sd) LINE-1 methylation was 65.3% (3.4) with a
relationships with epigenetic and asthma outcomes. range of 56.1–73.2%. Mean (sd) IFNγ CpG-54 was 79.6%
Analyses were adjusted for age (continuous) and gender. (4.5) with a range of 68.6–92.4%. Mean (sd) IFNγ CpG-
Although this diet and epigenetics repeated measures 186 was 70.1% (6.6) with a range of 49.1–81.6%. IFNγ
study was not directly related to the indoor air quality CpG-54 and IFNγ CpG-186 observations were moder-
intervention study, we included in our models indica- ately correlated with each other (r  =  0.42; p  =  0.02) as
tors for pre- versus post-intervention winter and home were IFNγ CpG-54 and LINE-1 methylation (r  =  0.44;
intervention assignment (i.e., placebo versus air filter). p = 0.01). IFNγ CpG-186 and LINE-1 methylation were
Inclusion of the following potential confounders: pres- not significantly correlated (r = 0.26; p = 0.15).
ence of cat or dog in home (yes or no), family income
(above or below $50,000) and parent education (col- Table 1  Selected characteristics of subset of ARTIS partici-
lege degree or no college degree) appreciably impacted pants included in epigenetic study
parameter estimates. Therefore, the final model included N Mean SD Range
age, gender, winter and intervention group assignment,
Age 32 12.8 2.5 8.0–17.0
presence of cat or dog, income and education. We inves-
Gender
tigated relationships between epigenetic markers and
 Male 15 (46%)
asthma measures in a similar manner. Due to the num-
 Female 17 (53%)
ber of comparisons (n = 264), a false discovery rate cor-
Ethnicity
rection [40] was applied and adjusted p-values (q-values)
 Non-Hispanic 30 (94%)
were calculated for each relationship where the GEE
BMI percentile 32 70.6 24.1 5.8–99.0
model was used. A threshold for significance was set at
dPFV 32 20.0 14.0 2.0–66.0
q < 0.20 which means we accept that 20% of the observed
% Predicted morning FEV1 32 81.7 19.7 26.9–112.9
significant relationships (i.e. 3.4 out of 17) could be false
% Predicted evening FEV1 32 82.3 19.3 19.3–110.1
positives.
PAQLQ 32 5.6 1.1 3.1–7.0
LINE-1 (%) 32 65.3 3.4 56.1–73.2
Results
A subset of 32 children from the ARTIS cohort partici- IFNγ-54 (%) 32 79.6 4.5 68.6–92.4
pated in this study of diet, asthma health and epigenet- IFNγ-186 (%) 32 70.1 6.6 49.1–81.6
ics and was included in the analyses described here. Diet SD standard deviation, BMI body mass index, dPFV evening to morning peak
flow variability, FEV1 forced expiratory volume in 1 s, PAQLQ pediatric asthma
data was collected once per winter in conjunction with quality of life questionnaire, IFNγ interferon gamma
Montrose et al. Allergy Asthma Clin Immunol (2017) 13:12
Evaluating dietary nutrients with respect to respiratory
health
When considered across categories of calculated intake,
most dietary nutrients failed to show a consistent asso-
ciation with respiratory health measures, but several
differences in PAQLQ scores were observed between
participants in the highest third versus the lowest third of
intake for some nutrients (Table 2). Phosphatidylcholine
was the only selected nutrient that was associated with
any of the three pulmonary function measure assessed.
Children in the middle tertile relative to the lowest had
16.04% point (95% CI 3.31, 28.78; q = 0.16) higher % pre-
dicted evening FEV1. Intake of selenium and folate was
associated with better PAQLQ scores. Specifically, par-
ticipants with the highest tertile of selenium and folate
intake had 1.4 unit (95% CI 0.90, 1.91; q = 0. 01) and 0.92
unit (95% CI 0.31, 1.53; q = 0.11) higher PAQLQ scores,
respectively. Additionally, nutrients in the one-carbon
metabolism cycle, phosphocholine (1.11 unit higher
PAQLQ score; 95% CI 0.23, 1.98; q  =  0.16) and betaine
(0.98 unit higher PAQLQ score; 95% CI 0.30, 1.66; q = 0.
13) were positively associated with PAQLQ.
Evaluating dietary nutrients with respect to methylation
outcomes
Intake of several nutrients was associated with LINE-1
methylation and methylation at CpG promoter site
IFNγ-186, but not for IFNγ-54 (Table 3). Children in the
highest tertile of kilocalories (3.2% points higher meth-
ylation; 95% CI 0.82, 5.58; q = 0.16) or the middle tertile
of protein (2.67% points higher methylation; 95% CI 0.62,
4.71; q = 0.16) had higher LINE-1 methylation. Similarly,
those in the highest tertile of methyl donating nutrients
free choline (2.18% points higher methylation; 95% CI
0.54, 3.82; q  =  0.16), total choline (2.60% points higher
methylation; 95% CI 0.60, 4.60; q  =  0.16) and folate
(4.29% points higher methylation; 95% CI 2.25, 6.34;
q  =  0.01) also had higher LINE-1 methylation. Intake
within the middle tertile of kilocalories (4.56% points
lower methylation; 95% CI −7.44, −1.69; q  =  0.09) and
folate (4.05% points lower methylation; 95% CI −6.18,
−1.19; q  =  0.02) compared to the lowest was associ-
ated with less IFNγ-186 methylation. However, intake
within middle tertile of monosaturated fat intake (6.88%
points higher methylation; 95% CI 3.11, 10.62; q = 0.02)
was associated with more IFNγ-186 methylation. Chil-
dren in the highest tertile of betaine intake (4.34% points
lower methylation; 95% CI −7.25, −1.42; q  =  0.12) and
both the middle and highest tertile of vitamin B6 intake
(6.57% points lower methylation; 95% CI −10.65, −2.48;
q  =  0.09 and 6.63% points lower methylation; 95% CI
−11.13, −2.14; q  =  0.12, respectively) had less IFNγ
CpG-186 methylation.
Page 5 of 12
Evaluating DNA methylation with respect to respiratory
health
We investigated the relationships between methylation
markers and asthma-related respiratory health meas-
urements (Table  4). A one-percentage point increase of
LINE-1 methylation was associated with a 1.24 percent-
age point (95% CI 0.31, 2.16; q = 0.16) increase in dPFV.
Neither IFNγ CpG-54 nor-186 methylation was associ-
ated with the respiratory health measures evaluated in
this study.
Discussion
In this study of dietary nutrients, DNA methylation and
asthma-related respiratory health outcomes, we observed
positive associations between several nutrients related to
one-carbon metabolism (e.g. folate, phosphocholine and
betaine) and the PAQLQ score. These nutrients were not
similarly associated with better self-monitored spirom-
etry outcomes, suggesting that the nutrients may posi-
tively influence asthma quality of life through another
mechanism. The composite PAQLQ score is comprised
of symptom, activity and emotion domains; however, a
post hoc analysis, which substituted individual domains
for the composite PAQLQ in a model with dietary nutri-
ents (i.e. those nutrients that had significant relationships
with composite PAQLQ), revealed no difference for the
impact of dietary intake on individual domains. Further-
more, the individual domains were highly correlated with
one another (data not shown) and therefore it is unclear
which of these domains may be more affected by diet.
Folate (or folic acid) is one of the most prominently
studied methyl donors and is known to play a role in
allergic asthma [42]. Many studies have investigated the
efficacy of methyl donating nutrient supplementation
to reduce the risk of asthma development. To date the
results have been inconclusive (see reviews [2, 43]). Many
of these supplementation studies have been either lim-
ited by ethical concerns or underpowered. We observed
that folate intake was associated with a higher PAQLQ
score in children with asthma. Based on the most recent
reviews, the evidence suggests that early life exposure to
folate has no major effects on asthma outcomes later in
life [44]. However, relevant to our findings, few studies
have investigated the relationship between folate intake
and asthma measures in adolescent children with estab-
lished asthma [45]. Folate deficiency has been associated
with asthma-related symptoms and exacerbations [46–
48]. Folate and betaine are involved in DNA methyla-
tion through the formation of S-adenosylmethionine and
homocysteine metabolism and therefore have the poten-
tial to affect gene expression, thereby influencing asthma
pathogenesis [43]. In addition to methyl donors, selenium
was also positively associated with PAQLQ score. Fabian
Montrose et al. Allergy Asthma Clin Immunol (2017) 13:12 Page 6 of 12

Table 2  The relationship between  select nutrients and  asthma health measures by  dietary tertiles in  ARTIS where  the
lowest tertile of intake (T1) is the reference group
Asthma health dPFV % Predicted morning FEV1 % Predicted evening FEV1 PAQLQ
measure
Dietary factor Estimate (95% CI) q-value Estimate (95% CI) q-value Estimate (95% CI) q-value Estimate (95% CI) q-value

Kilocalories
 T3:T1 −3.18 (−11.41, 5.06) 0.80 13.09 (−4.04, 30.22) 0.61 4.18 (−14.08, 22.44) 0.87 0.62 (−0.25, 1.48) 0.68
 T2:T1 −0.50 (−8.78, 7.78) 0.98 0.83 (−13.94, 15.60) 0.98 7.04 (−8.31, 22.40) 0.76 −0.34 (−1.20, 0.51) 0.79
Protein
 T3:T1 −6.96 (−20.40, 6.48) 0.72 19.26 (0.48, 38.05) 0.33 10.96 (−2.58, 24.51) 0.55 0.70 (−0.16, 1.56) 0.55
 T2:T1 −3.42 (−14.56, 7.73) 0.83 6.31 (−8.30, 20.92) 0.77 1.14 (−12.42, 14.70) 0.97 0.59 (−0.09, 1.27) 0.51
M-fats
 T3:T1 −2.87 (−10.62, 4.87) 0.81 11.64 (−6.82, 30.11) 0.71 8.55 (−10.72, 27.81) 0.77 0.39 (−0.31, 1.09) 0.71
 T2:T1 −0.57 (−8.81, 7.67) 0.98 3.89 (−9.58, 17.36) 0.83 3.26 (−9.25, 15.77) 0.85 0.52 (−0.03, 1.07) 0.41
S-fats
 T3:T1 −3.99 (−15.87, 7.89) 0.83 16.85 (1.36, 32.35) 0.29 11.73 (0.65, 22.82) 0.33 −0.16 (−0.82, 0.50) 0.87
 T2:T1 −6.56 (−16.72, 3.59) 0.70 16.67 (2.96, 30.37) 0.28 10.19 (−0.41, 20.80) 0.41 −0.05 (−0.77, 0.67) 0.98
Omega 3:6 ratio
 T3:T1 −1.25 (−10.54, 8.03) 0.92 2.42 (−11.59, 16.44) 0.90 −0.39 (−13.84, 13.07) 1.00 −0.13 (−1.18, 0.93) 0.92
 T2:T1 −1.29 (−12.09, 9.51) 0.92 5.40 (−4.00, 14.80) 0.71 1.46 (−6.18, 9.10) 0.90 0.11 (−0.54, 0.75) 0.90
Selenium
 T3:T1 5.44 (−2.98, 13.86) 0.70 6.63 (−6.31, 19.56) 0.73 6.37 (−6.57, 19.30) 0.74 1.40 (0.90, 1.91) 0.01*
 T2:T1 2.52 (−4.56, 9.59) 0.82 0.20 (−14.11, 14.51) 1.00 2.91 (−12.96, 18.77) 0.90 −0.56 (−1.59, 0.47) 0.71
Fiber
 T3:T1 −0.39 (−9.02, 8.25) 0.99 7.75 (−5.27, 20.78) 0.71 5.86 (−8.02, 19.74) 0.77 0.66 (−0.52, 1.84) 0.71
 T2:T1 0.63 (−9.85, 11.11) 0.98 −5.36 (−17.98, 7.27) 0.77 −3.60 (−15.79, 8.59) 0.83 0.32 (−0.38, 1.02) 0.76
Folate
 T3:T1 −3.43 (−14.76, 7.89) 0.83 10.07 (−7.95, 28.09) 0.71 3.75 (−10.53, 18.02) 0.85 0.92 (0.31, 1.53) 0.11*
 T2:T1 −3.03 (−14.23, 8.17) 0.85 −0.03 (−17.50, 17.44) 1.00 −0.24 (−13.99, 13.50) 1.00 −0.25 (−0.83, 0.35) 0.78
Methionine
 T3:T1 −4.83 (−13.90, 4.24) 0.72 14.81 (−0.25, 29.88) 0.36 11.07 (−4.02, 26.17) 0.65 0.52 (−0.37, 1.41) 0.71
 T2:T1 1.87 (−7.53, 11.28) 0.90 5.22 (−12.12, 22.57) 0.83 7.29 (−10.16, 24.73) 0.77 −0.06 (−1.58, 1.46) 0.99
Free choline
 T3:T1 2.44 (−4.68, 9.56) 0.83 11.20 (−6.45, 28.86) 0.70 7.69 (−10.98, 26.36) 0.78 0.00 (−1.34, 1.34) 1.00
 T2:T1 5.35 (−3.47, 14.17) 0.71 −2.64 (14.74, 9.47) 0.88 −0.21 (−13.39, 12.97) 1.00 −0.31 (−1.26, 0.64) 0.83
Glycpp-choline
 T3:T1 2.84 (−5.75, 11.42) 0.83 −3.05 (−18.17, 12.07) 0.89 −3.64 (−19.88, 12.61) 0.88 0.61 (−0.17, 1.39) 0.59
 T2:T1 −1.81 (−9.22, 5.59) 0.87 2.88 (−9.04, 14.80) 0.87 3.70 (−8.10, 15.50) 0.83 0.12 (−0.41, 0.65) 0.87
Pp-choline
 T3:T1 4.54 (−3.07, 12.15) 0.71 −2.95 (−12.99, 7.08) 0.83 −5.21 (−15.57, 5.14) 0.73 1.11 (0.23, 1.98) 0.16*
 T2:T1 −1.03 (−7.55, 5.50) 0.90 −10.81 (−27.27, 5.64) 0.69 −12.35 (−29.79, 5.09) 0.68 0.50 (−0.61, 1.60) 0.77
Ppt-choline
 T3:T1 0.36 (−6.86, 7.58) 0.98 11.60 (−1.45, 24.65) 0.48 −1.13 (−19.21, 16.94) 0.98 0.24 (−0.86, 1.35) 0.88
 T2:T1 5.71 (−2.92, 14.33) 0.69 −3.17 (−15.51, 9.17) 0.85 16.04 (3.31, 28.78) 0.16* −0.19 (−0.58, 0.21) 0.76
Total choline
 T3:T1 1.33 (−6.50, 9.16) 0.90 8.61 (−7.92, 25.14) 0.72 5.96 (−12.23, 24.15) 0.83 0.18 (−1.26, 1.61) 0.92
 T2:T1 3.76 (−4.42, 11.94) 0.76 −6.68 (−20.25, 6.90) 0.74 −2.06 (−16.74, 12.62) 0.92 −0.38 (−1.35, 0.59) 0.80
Betaine
 T3:T1 −2.62 (−9.35, 4.10) 0.80 12.85 (0.98, 24.72) 0.29 9.13 (−1.59, 19.85) 0.51 0.98 (0.30, 1.66) 0.13*
 T2:T1 4.20 (−6.18, 14.58) 0.79 −9.34 (−21.55, 2.86) 0.61 −9.10 (−19.78, 3.59) 0.68 0.21 (−0.56, 0.98) 0.85
Vitamin B-2
 T3:T1 −0.95 (−8.43, 6.54) 0.92 −4.74 (−18.10, 8.61) 0.82 3.81 (−10.97, 18.58) 0.85 0.40 (−0.66, 1.46) 0.80
 T2:T1 1.66 (−8.13, 11.44) 0.90 8.98 (−7.57, 25.52) 0.71 −4.04 (−17.65, 9.56) 0.83 −0.12 (1.12, 0.88) 0.92
Montrose et al. Allergy Asthma Clin Immunol (2017) 13:12
Table 2  continued
Asthma health dPFV % Predicted morning FEV1
measure
Dietary factor Estimate (95% CI) q-value Estimate (95% CI)
Vitamin B-6
 T3:T1 2.53 (−5.37, 10.43) 0.83 4.19 (−10.21, 18.59)
 T2:T1 4.64 (−3.10, 12.38) 0.71 −4.96 (−20.50, 10.57)
Vitamin B-12
 T3:T1 −0.02 (−8.26, 8.22) 1.00 6.42 (−8.93, 21.77)
 T2:T1 1.53 (−7.17, 10.23) 0.90 −2.40 (−16.46, 11.66)
M-fats monosaturated fats, S-fats saturated fats, Glycpp-choline glycerophosphocholine, Pp-choline phosphocholine, Ppt-choline phosphotidylcholine, dPFV evening to
morning peak flow variability, FEV1 forced expiratory volume in 1 s, PAQLQ pediatric asthma quality of life questionnaire
* q-value < 0.20
et  al. found that children with asthma compared to
healthy control children had lower plasma levels of sele-
nium and higher exhaled nitric oxide, a marker of poor
lung health [49]. However, a group of Swedish research-
ers found no impact of selenium intake on allergic disease
in young children [50]. The inconclusive results regarding
the impact of selenium intake on allergic asthma could
be attributed to the fact that while selenium does have
antioxidant properties it also has the ability to upregulate
some immune responses [17, 51]. In our analysis sele-
nium status is associated with better asthma quality of
life measures, however, this nutrient was not associated
with LINE-1 or IFNγ methylation profiles.
Among the dietary nutrients investigated in this study
only phosphatidylcholine was modestly associated
with self-administered spirometry measures, specifi-
cally higher evening FEV1. Phosphatidylcholine is phos-
pholipid and a major dietary source of choline, which is
involved in one-carbon metabolism. Phospholipids can
also impact T cell function in a number of ways includ-
ing membrane fluidity and gene expression, which could
have indirect immunomodulatory effects [52]. There-
fore, our observation of a positive association between
phosphatidylcholine and FEV1 in these children could
be reflective of reduced lung inflammation. However,
this association was not consistent across tertiles of
phosphatidylcholine intake nor was there a consistent
response across the different FEV1 measures, evening
versus morning. These concerns together with the stud-
ies that have linked phosphatidylcholine to cardiovascu-
lar disease [53] and related inflammatory symptoms [54],
suggest that the association between phosphatidylcholine
intake and FEV1 should be interpreted with caution.
In addition to evaluating the relationship between
nutritional intake and asthma health, a potential epige-
netic mechanism was examined by determining if global
methylation or IFNγ promoter methylation was associ-
ated with nutritional intake. Global methylation, though
Page 7 of 12
% Predicted evening FEV1 PAQLQ
q-value Estimate (95% CI) q-value Estimate (95% CI) q-value
0.83 0.26 (−13.43, 13.94) 1.00 0.46 (−0.21, 1.13) 0.69
0.83 0.81 (−14.85, 16.46) 0.98 −0.65 (−1.20, 0.00) 0.36
0.77 1.70 (−11.74, 15.13) 0.92 0.79 (0.09, 1.50) 0.29
0.90 −3.38 (−16.85, 10.10) 0.86 0.29 (−0.53, 1.11) 0.82
informative, is difficult to interpret in the context of res-
piratory health and may be even more complex in this
cohort of children with asthma. Few studies have looked
at buccal DNA LINE-1 global methylation in healthy
children. A study of 57 healthy girls aged from 6 to 15
investigated LINE-1 global methylation in saliva samples
and the average (SD) was 75.2 (3.4) [55]. The cells col-
lected from buccal and saliva should be similar, however
the mean LINE-1 methylation in our study overall was
considerably lower, which could be attributed to asthma
status. It is clear from the literature that intake of methyl
donors can result in measureable changes to the mam-
malian epigenome [20]. Our study shows that dietary
intake of folate, free choline, and total choline is posi-
tively associated with LINE-1 methylation. By conven-
tion, an increase in global methylation is thought to be
protective, while a shift toward genome-wide hypometh-
ylation is often associated with a poor health outcome
or disease [56, 57]. Nevertheless, our data suggested that
global methylation was positively associated with dPFV,
an indicator of airway hyper-reactivity. This finding
could be a characteristic of the study population, which
had relatively low average global methylation. Further,
DNA methylation is dynamic and global methylation is
a reflection of the epigenetic changes occurring at many
gene locations.
Our study specifically focused on IFNγ as a candidate
gene, hypothesizing that this gene would lie in the mech-
anistic pathway linking dietary intake to asthma health
outcomes in children. Previous studies have established
that IFNγ CpG-54 and -186 (-53 and -190 are the cor-
responding murine CpGs) are relevant to allergic out-
comes in animal models [58, 59] and humans [60–62].
In the mouse, Jones et  al. showed that these CpGs are
functionally relevant (i.e., methylation status affects tran-
scription of the IFNγ gene) and that de novo methyla-
tion of these sites plays a key role in Th2 polarization at
least within the CD4+ T cells [63]. In an human asthma
Montrose et al. Allergy Asthma Clin Immunol (2017) 13:12 Page 8 of 12

Table 3  The relationship between select nutrients and DNA methylation markers by dietary tertiles in ARTIS, where the
lowest tertile of intake (T1) is the reference group
Epigenetic measure IFNγ CpG-54 IFNγ CpG-186 LINE-1
Dietary factor Estimate (95% CI) q-value Estimate (95% CI) q-value Estimate (95% CI) q-value

Kilocalories
 T3:T1 0.46 (−2.46, 3.38) 0.90 −0.11 (−4.35, 4.13) 1.00 3.20 (0.82, 5.58) 0.16*
 T2:T1 −0.40 (−2.87, 2.07) 0.90 −4.56 (−7.44, −1.69) 0.09* 0.00 (−2.45, 2.46) 1.00
Protein
 T3:T1 2.60 (−0.90, 6.10) 0.65 −0.04 (−3.41, 3.32) 1.00 1.62 (−0.79, 4.03) 0.69
 T2:T1 2.17 (−0.46, 4.80) 0.55 2.02 (−1.24, 5.27) 0.71 2.67 (0.62, 4.71) 0.16*
M-fats
 T3:T1 −1.10 (−3.70, 1.51) 0.77 4.77 (−0.04, 9.57) 0.36 1.14 (−1.23, 3.51) 0.76
 T2:T1 −2.42 (−4.74, −0.11) 0.33 6.86 (3.11, 10.62) 0.02* 1.12 (−1.00, 3.24) 0.72
S-fats
T3:T1 −0.75 (−3.11, 1.61) 0.83 5.11 (−1.08, 11.31) 0.55 1.52 (−1.09, 4.13) 0.71
T2:T1 1.02 (−1.91, 3.96) 0.82 3.00 (−2.48, 8.48) 0.71 1.26 (−1.45, 3.97) 0.76
Omega 3:6 ratio
 T3:T1 1.44 (−1.23, 4.10) 0.71 2.29 (−1.85, 6.44) 0.71 −0.09 (−2.70, 2.52) 0.99
 T2:T1 3.72 (0.44, 7.01) 0.29 0.24 (−4.72, 5.20) 0.98 0.96 (−1.03, 2.95) 0.74
Selenium
 T3:T1 0.34 (−2.45, 3.13) 0.92 −1.40 (−5.17, 2.38) 0.81 2.32 (−0.15, 4.79) 0.45
 T2:T1 1.40 (−0.60, 3.40) 0.68 −2.38 (−5.01, 0.24) 0.45 −0.18 (−2.69, 2.33) 0.98
Fiber
 T3:T1 0.53 (−2.31, 3.37) 0.90 −4.67 (−9.09, −0.24) 0.33 1.89 (−1.02, 4.80) 0.69
 T2:T1 0.19 (−2.29, 2.68) 0.98 −2.52 (−6.54, 1.49) 0.71 −0.27 (−2.70, 2.16) 0.94
Folate
 T3:T1 2.93 (0.06, 5.80) 0.36 2.46 (−2.25, 7.17) 0.72 4.29 (2.25, 6.34) 0.01*
 T2:T1 1.95 (−0.03, 3.93) 0.36 −4.05 (−6.18, −1.91) 0.02* 0.46 (−1.24, 2.16) 0.85
Methionine
 T3:T1 0.63 (−2.42, 3.68) 0.89 −2.70 (−6.86, 1.45) 0.69 1.12 (−1.78, 4.03) 0.80
 T2:T1 0.43 (−2.25, 3.11) 0.90 −1.38 (−5.68, 2.92) 0.83 1.58 (−1.04, 4.21) 0.71
Free choline
 T3:T1 1.33 (−1.08, 3.74) 0.71 −2.61 (−5.84, 0.63) 0.55 2.18 (0.54, 3.82) 0.16*
 T2:T1 1.29 (−1.03, 3.60) 0.71 −3.28 (−6.33, −0.24) 0.29 0.01 (−1.82, 1.85) 1.00
Glycpp-choline
 T3:T1 0.44 (−2.23, 3.12) 0.90 −2.81 (−5.95, 0.33) 0.48 0.66 (−1.48, 2.80) 0.83
 T2:T1 1.68 (−0.54, 3.89) 0.65 −0.31 (−3.84, 3.22) 0.97 0.24 (−1.74, 2.21) 0.92
Pp-choline
 T3:T1 0.50 (−2.65, 3.65) 0.90 −1.77 (−5.27, 1.73) 0.73 0.41 (−1.43, 2.25) 0.88
 T2:T1 1.47 (−0.52, 3.46) 0.65 −1.54 (−5.83, 2.75) 0.82 −2.08 (−4.41, 0.24) 0.48
Ppt-choline
 T3:T1 1.49 (−1.08, 4.06) 0.71 −2.33 (−6.06, 1.39) 0.71 0.64 (−1.66, 2.94) 0.84
 T2:T1 −0.78 (−3.40, 1.84) 0.83 −1.55 (−6.86, 3.76) 0.83 −1.66 (−4.03, 0.71) 0.68
Total choline
 T3:T1 1.29 (−1.06, 3.64) 0.71 −3.47 (−6.68, −0.26) 0.29 2.60 (0.60, 4.60) 0.16*
 T2:T1 0.85 (−1.41, 3.11) 0.80 −3.58 (−6.59, −0.58) 0.28 0.42 (−1.88, 2.73) 0.90
Betaine
 T3:T1 −0.00 (−2.50, 2.49) 1.00 −4.34 (−7.25, −1.42) 0.12* 1.18 (−1.25, 3.62) 0.74
 T2:T1 1.52 (−1.21, 4.26) 0.71 1.28 (−4.28, 1.73) 0.77 1.65 (−0.61, 3.91) 0.65
Vitamin B-2
 T3:T1 0.62 (−2.44, 3.68) 0.89 −4.19 (−7.91, −0.48) 0.29 2.47 (−0.37, 5.32) 0.51
Montrose et al. Allergy Asthma Clin Immunol (2017) 13:12 Page 9 of 12

Table 3  continued
Epigenetic measure IFNγ CpG-54 IFNγ CpG-186 LINE-1
Dietary factor Estimate (95% CI) q-value Estimate (95% CI) q-value Estimate (95% CI) q-value

 T2:T1 1.41 (−0.74, 3.56) 0.69 −4.03 (−7.86, −0.20) 0.33 0.23 (−2.48, 2.94) 0.97
Vitamin B-6
 T3:T1 2.05 (−1.04, 5.14) 0.69 −6.63 (−11.13, −2.14) 0.12* 1.44 (−1.65, 4.52) 0.76
 T2:T1 0.66 (−1.80, 3.12) 0.85 −6.57 (−10.65, −2.48) 0.09* 0.13 (−2.39, 2.65) 0.98
Vitamin B-12
 T3:T1 1.12 (−2.31, 4.55) 0.83 −2.31 (−7.00, 2.39) 0.74 0.59 (−2.43, 3.60) 0.90
 T2:T1 2.21 (−0.48, 4.91) 0.55 1.84 (−2.37, 6.05) 0.77 0.87 (−2.09, 3.82) 0.83
M-fats monosaturated fats, S-fats saturated fats, Glycpp-choline glycerophosphocholine, Pp-choline phosphocholine, Ppt-choline phosphotidylcholine, IFNγ interferon
gamma
* q-value <0.20

Table 4  The relationship between epigenetic measurements and asthma health outcomes in ARTIS

Epigenetic marker dPFV % Predicted morning FEV1 % Predicted evening FEV1 PAQLQ
Estimate (95% CI) q-value Estimate (95% CI) q-value Estimate (95% CI) q-value Estimate (95% CI) q-value

LINE-1 1.24 (0.31, 2.16) 0.16* −0.89 (−3.18, 1.41) 0.80 −1.19 (−3.41, 1.03) 0.72 0.04 (−0.11, 0.20) 0.84
IFNγ CpG-186 0.58 (0.06, 1.09) 0.29 −0.53 (−1.79, 0.72) 0.77 −0.57 (−1.62, 0.48) 0.72 0.03 (0.00, 0.07) 0.44
IFNγ CpG-54 0.22 (−0.74, 1.19) 0.87 −0.57 (−1.92, 0.77) 0.77 −0.84 (−2.10, 0.42) 0.69 0.02 (−0.07, 0.11) 0.85
dPFV evening to morning peak flow variability, FEV1 forced expiratory volume in 1 s, PAQLQ pediatric asthma quality of life questionnaire, IFNγ interferon gamma
* q-value <0.20

cohort, Lovinsky-Desir et al. showed that there are differ-
ential methylation profiles for these CpGs relative to age,
sex and tissue type [64]. For example, when methylation
profiles of buccal cells and CD4+  lymphocytes isolated
from whole blood were compared, IFNγ CpG-186 was
correlated for males but not females. Further, methyla-
tion values for IFNγ CpG-54 and -186 were correlated for
children and adults in CD4+  lymphocytes but only for
adults in buccal cells. White et al. investigated IFNγ pro-
moter methylation profiles by in vitro polyclonal expan-
sion of CD4+ and CD8+ T cells sorted from peripheral
blood mononuclear cells [62]. When samples collected
from adolescent children were stratified by atopic sta-
tus, the authors found that, for CD8+ T cells under Th1
polarizing conditions, IFNγ CpG sites -54 and -186 were
less methylated in the non-atopic children.
When evaluating the impact of diet on IFNγ promoter
methylation, we found that only IFNγ CpG-186 meth-
ylation patterns were affected by selected nutrients. We
observed that intake of kilocalories and three methyl
donating nutrients was associated with less IFNγ CpG-
186 methylation, while children who had higher intake
of monosaturated fats had more IFNγ CpG-186 meth-
ylation. Based on the functional data available for this
CpG site, which we note does not come from buccal
cells, we speculate that a negative association between
a dietary nutrient and methylation at this site could
impact the Th1/Th2 balance by increasing the expres-
sion of IFNγ. Overall nutrient intake has previously been
linked to IFNγ production [65]. However, overnutrition
is not likely a preferable or effective asthma interven-
tion especially due to the potential links between obe-
sity, inflammation, and asthma. Monosaturated fat was
the only nutrient we found to be positively associated
with IFNγ-186 methylation. In a study of approximately
1200 adolescent children conducted in Taiwan, intake
of monosaturated fats was inversely associated with risk
of asthma [66]. By contrast, a study of nearly 4000 adult
European participants found that intake of monosatu-
rated fats was positively associated with allergic sensiti-
zation [67].
IFNγ CpG promoter methylation at site -54 and -186
was not associated with respiratory health measures
or PAQLQ. This suggests that the positive relationship
revealed between PAQLQ composite score and intake of
selenium, folate, phosphocholine, and betaine may not
be working directly through epigenetic modification of
these specific sites as we had hypothesized.
Limitations and cautions
We evaluated several dietary macro- and micronutri-
ents in this study, but these factors likely include only a
Montrose et al. Allergy Asthma Clin Immunol (2017) 13:12
portion of the exogenous factors that could influence
DNA methylation in this population. While FFQs are
an accepted and validated method for acquiring per-
sonal dietary information, we note that the portion sizes
and specific foodstuffs were self-reported by the partici-
pants with assistance and input from parents. Though
it is widely accepted, BMI may actually be a poor indi-
cator for obesity in children and adolescents who have
large, lean body mass from physical activity, high mus-
cularity, or frame size. By focusing on select candidate
DNA methylation markers, we recognize that numerous
inflammatory pathways involving diet and asthma may
not have been captured. We also are limited in our inter-
pretation of the DNA methylation data because we did
not measure IFNγ expression or protein levels in these
samples. For example, we found some dietary factors to
be negatively associated with DNA methylation, which
could be informative for asthma intervention strategies,
but such interpretations require further assessment as
methylation changes do not necessarily translate to func-
tional changes in the target tissue. Finally, although we
accounted for false discoveries, we recognize that several
statistical tests were performed and would expect some
significant results due to chance alone. Thus these obser-
vations should be considered exploratory and requiring
of further study in other populations.
Conclusions
Within this cohort of childhood asthmatics, we sought
to identify dietary nutrients that may be beneficial for
respiratory health. In addition we measured LINE-1
and IFNγ (CpG-54 and -186) methylation levels to iden-
tify pathways whereby diet influences the health among
children with asthma. In this study, selenium and several
nutrients involved in the one-carbon metabolism path-
way were associated with improved asthma quality of
life measures. Furthermore, these data showed that some
dietary constituents were associated with both global and
gene specific methylation in children with asthma. The
two IFNγ CpG sites that were investigated appear to be
uniquely affected by intake of micro- and macronutrients.
Additional file
Additional file 1: Table S1. Primers and amplification conditions for
PCR and pyrosequencing experiments. Figure S1. Description of ARTIS
sub-study design.
Abbreviations
ARTIS: asthma randomized trial of indoor wood smoke; BMI: body mass
index; CDC: U.S. Centers for Disease Control and Prevention; CI: confidence
interval; CpG: regions of DNA where a cytosine nucleotide is followed by a
guanine nucleotide; DNA: deoxyribonucleic acid; dPFV: diurnal peak flow
variability; FEV1: forced expiratory volume in the first second; FFQ: food
frequency questionnaires; FOXp3: transcription factor forkhead box p3; GEE:
Page 10 of 12
generalized estimating equations; glycpp-choline: glycerophosphocholine;
IFNγ: interferon gamma; LINE-1: long interspersed nuclear elements; M-fats:
monosaturated fats; NO2: nitrogen dioxide; PAQLQ: pediatric asthma quality of
life questionnaire; PCR: polymerase chain reaction; PEF: peak expiratory flow;
PM: particulate matter; PM2.5: fine PM or particles less than 2.5 micrometers in
aerodynamic diameter; Pp-choline: phosphocholine; Ppt-choline: phosphoti-
dylcholine; Runx3: runt-related transcription factor 3; S-fats: saturated fats; SD:
standard deviation; Tregs: regulatory T cells; Th1: type 1 T helper; Th2: type 2 T
helper; USEPA: U.S. Environmental Protection Agency.
Authors’ contributions
LM aided in the design of the study, acquisition and analysis of data, and with
drafting the manuscript. TW aided in the design of the study and with drafting
the manuscript. ES contributed to the data analysis strategy and interpretation
as well as with drafting the manuscript. YHC provided support for epigenetic
assays and with manuscript revisions. BB aided in the design, collection, and
analysis of diet related data as well as manuscript revisions. CN aided in the
design of the study, data analysis strategy and interpretation, and with draft-
ing the manuscript. All authors read and approved the final manuscript.
Author details
1
 School of Public Health, University of Michigan, 1420 Washington Heights,
Ann Arbor, MI 48109, USA. 2 Center for Environmental Health Sciences, Uni-
versity of Montana, 32 Campus Drive‑159 Skaggs, Missoula, MT 59812, USA.
3
 Department of Health and Human Performance, University of Montana, 32
Campus Drive, Missoula, MT 59812, USA.
Acknowledgements
The authors would like to thank Dr. Jackie Goodrich and Virginia Kay for
providing support for epigenetic assay development, Emily Weiler for field col-
lection of data, Dr. Dana Dolinoy for mentorship and Carolyn Hester for aiding
in study management and organization.
Competing interests
The authors declare that they have no competing interests.
Availability of data and materials
Please contact author for data requests.
Ethics approval and consent to participate
The study was approved by the University of Montana Institutional Review
Board (protocol numbers 233_10 and 152_11). In addition to the informed
consent procedures for the parent study, children were separately assented
to participate in this diet and epigenetics study and parents signed a parental
permission and informed consent form.
Funding
This research was supported by the National Institute of Environmental Health
Sciences (1R01ES016336–01 and 3R01ES016336–02S1). Additional support was
provided by a COBRE award (P20RR017670) and the National Institute of Gen-
eral Medical Sciences (P20GM103474 and 1U54GM104944). LM was partially
supported by an NIH NRSA T32 training program Grant (#E2007062, PI: Rita
Loch-Caruso). The content is solely the responsibility of the authors and does
not necessarily represent the official views of the National Institutes of Health.
Received: 8 June 2016 Accepted: 17 February 2017
References
1. Akinbami LJ, et al. Trends in asthma prevalence, health care use, and
mortality in the United States, 2001–2010. NCHS Data Brief. 2012;94:1–8.
2. Han YY, et al. Diet and asthma: vitamins and methyl donors. Lancet Respir
Med. 2013;1(10):813–22.
3. Thuesen BH, et al. Atopy, asthma, and lung function in relation to folate
and vitamin B(12) in adults. Allergy. 2010;65(11):1446–54.
4. Mehta AK, et al. Choline attenuates immune inflammation and sup-
presses oxidative stress in patients with asthma. Immunobiology.
2010;215(7):527–34.
Montrose et al. Allergy Asthma Clin Immunol (2017) 13:12
5. Barros R, et al. Adherence to the Mediterranean diet and fresh fruit intake
are associated with improved asthma control. Allergy. 2008;63(7):917–23.
6. Barros R, et al. Dietary intake of alpha-linolenic acid and low ratio of
n-6:n-3 PUFA are associated with decreased exhaled NO and improved
asthma control. Br J Nutr. 2011;106(3):441–50.
7. Castro-Rodriguez JA, et al. Mediterranean diet as a protective factor for
wheezing in preschool children. J Pediatr. 2008;152(6):823–8.
8. Chatzi L, et al. Diet, wheeze, and atopy in school children in Menorca,
Spain. Pediatr Allergy Immunol. 2007;18(6):480–5.
9. Chatzi L, et al. Mediterranean diet in pregnancy is protective for wheeze
and atopy in childhood. Thorax. 2008;63(6):507–13.
10. de Batlle J, et al. Mediterranean diet is associated with reduced asthma
and rhinitis in Mexican children. Allergy. 2008;63(10):1310–6.
11. Garcia-Marcos L, et al. Relationship of asthma and rhinoconjunctivitis
with obesity, exercise and Mediterranean diet in Spanish schoolchildren.
Thorax. 2007;62(6):503–8.
12. Romieu I, et al. Dietary intake, lung function and airway inflammation
in Mexico City school children exposed to air pollutants. Respir Res.
2009;10:122.
13. Berthon BS, Wood LG. Nutrition and respiratory health-feature review.
Nutrients. 2015;7(3):1618–43.
14. Berthon BS, et al. Investigation of the association between dietary
intake, disease severity and airway inflammation in asthma. Respirology.
2013;18(3):447–54.
15. Cortessis VK, et al. Environmental epigenetics: prospects for studying
epigenetic mediation of exposure–response relationships. Hum Genet.
2012;131(10):1565–89.
16. Lovinsky-Desir S, Miller RL. Epigenetics, asthma, and allergic dis-
eases: a review of the latest advancements. Curr Allergy Asthma Rep.
2012;12(3):211–20.
17. Norton RL, Hoffmann PR. Selenium and asthma. Mol Aspects Med.
2012;33(1):98–106.
18. Bansal P, et al. Intranasal administration of a combination of choline chlo-
ride, vitamin C, and selenium attenuates the allergic effect in a mouse
model of airway disease. Free Radic Biol Med. 2014;73:358–65.
19. North ML, Ellis AK. The role of epigenetics in the developmental origins of
allergic disease. Ann Allergy Asthma Immunol. 2011;106(5):355–61.
20. Dolinoy DC. The agouti mouse model: an epigenetic biosensor for nutri-
tional and environmental alterations on the fetal epigenome. Nutr Rev.
2008;66(Suppl 1):S7–11.
21. Hollingsworth JW, et al. In utero supplementation with methyl
donors enhances allergic airway disease in mice. J Clin Invest.
2008;118(10):3462–9.
22. Lloyd CM, Hawrylowicz CM. Regulatory T cells in asthma. Immunity.
2009;31(3):438–49.
23. Nadeau K, et al. Ambient air pollution impairs regulatory T-cell function in
asthma. J Allergy Clin Immunol. 2010;126(4):845–52.
24. Friso S, et al. One-carbon metabolism and epigenetics. Mol Aspects Med.
2016.
25. Calderon C, et al. T-cell cytokine profiles are altered in childhood asthma
exacerbation. Respirology. 2009;14(2):264–9.
26. Crinnion WJ. Do environmental toxicants contribute to allergy and
asthma? Altern Med Rev. 2012;17(1):6–18.
27. Smart JM, et al. Polyclonal and allergen-induced cytokine responses in
adults with asthma: resolution of asthma is associated with normalization
of IFN-gamma responses. J Allergy Clin Immunol. 2002;110(3):450–6.
28. Meng H, et al. In vitro production of IL-6 and IFN-gamma is influenced by
dietary variables and predicts upper respiratory tract infection incidence
and severity respectively in young adults. Front Immunol. 2015;6:94.
29. Mohammadi-Shahrokhi V, et al. Immunomodulatory effects of adjuvants
CPG, MPLA, and BCG on the Derp2-induced acute asthma at early life in
an animal model of BALB/c Mice. Inflammation. 2016.
30. Santeliz JV, et al. Amb a 1-linked CpG oligodeoxynucleotides reverse
established airway hyperresponsiveness in a murine model of asthma. J
Allergy Clin Immunol. 2002;109(3):455–62.
31. Noonan CW, Ward TJ. Asthma randomized trial of indoor wood
smoke (ARTIS): rationale and methods. Contemp Clin Trials.
2012;33(5):1080–7.
32. Ward TJ, et al. Efficacy of interventions targeting household air pol-
lution from residential wood stoves. J Expo Sci Environ Epidemiol.
2017;27(1):64–71.
Page 11 of 12
33. Semmens EO, et al. Indoor particulate matter in rural, wood stove heated
homes. Environ Res. 2015;138:93–100.
34. BMI percentile calculator for child and teen. https://nccd.cdc.gov/
dnpabmi/calculator.aspx. Accessed 1 Dec 2016.
35. Block G, et al. Revision of dietary analysis software for the health habits
and history questionnaire. Am J Epidemiol. 1994;139(12):1190–6.
36. Block G, DiSogra C. WIC dietary assessment validation study. Final report.
Alexandria: US Department of Agriculture, Food and Nutrition Service;
1995.
37. Block G, et al. Comparison of two dietary questionnaires validated against
multiple dietary records collected during a 1-year period. J Am Diet
Assoc. 1992;92:686–93.
38. Block G, et al. Validation of a self-administered diet history questionnaire
using multiple diet records. J Clin Epidemiol. 1990;43(12):1327–35.
39. Juniper EF, et al. Measuring quality of life in children with asthma. Qual
Life Res. 1996;5(1):35–46.
40. Benjamini Y, Hochberg Y. Controlling the false discovery rate—a
practical and powerful approach to multiple testing. J R Stat Soc Series
B-Methodol. 1995;57(1):289–300.
41. Lung National Heart, Institute Blood. Expert panel report 3: guidelines
for the diagnosis and management of asthma. Bethesda: National Heart
Lung, and Blood Institute; 2007.
42. Sharma S, Litonjua A. Asthma, allergy, and responses to methyl donor
supplements and nutrients. J Allergy Clin Immunol. 2014;133(5):1246–54.
43. Blatter J, et al. Folate and asthma. Am J Respir Crit Care Med.
2013;188(1):12–7.
44. Crider KS, et al. Prenatal folic acid and risk of asthma in children: a system-
atic review and meta-analysis. Am J Clin Nutr. 2013;98(5):1272–81.
45. Lin JH, et al. Relationships between folate and inflammatory features of
asthma. J Allergy Clin Immunol. 2013;131(3):918–20.
46. Matsui EC, Matsui W. Higher serum folate levels are associated
with a lower risk of atopy and wheeze. J Allergy Clin Immunol.
2009;123(6):1253–9.
47. Farres MN, et al. Study of folate status among Egyptian asthmatics. Intern
Med. 2011;50(3):205–11.
48. Blatter J, et al. Folate deficiency, atopy, and severe asthma exacerbations
in Puerto Rican children. Ann Am Thorac Soc. 2016;13(2):223–30.
49. Fabian E, et al. Nutritional supplements and plasma antioxidants in child-
hood asthma. Wien Klin Wochenschr. 2013;125(11–12):309–15.
50. Rosenlund H, et al. Antioxidant intake and allergic disease in children. Clin
Exp Allergy. 2012;42(10):1491–500.
51. Hoffmann PR. Selenium and asthma: a complex relationship. Allergy.
2008;63(7):854–6.
52. Pollock AH, et al. Prolonged Intake of dietary lipids alters mem-
brane structure and T cell responses in LDLr−/− Mice. J Immunol.
2016;196(10):3993–4002.
53. Wang Z, et al. Gut flora metabolism of phosphatidylcholine promotes
cardiovascular disease. Nature. 2011;472(7341):57–63.
54. de Pablo MA, Puertollano MA, Alvarez de Cienfuegos G. Biological and
clinical significance of lipids as modulators of immune system functions.
Clin Diagn Lab Immunol. 2002;9(5):945–50.
55. Wu HC, et al. Correlation of DNA methylation levels in blood and
saliva DNA in young girls of the LEGACY Girls study. Epigenetics.
2014;9(7):929–33.
56. Rozek LS, et al. Epigenetics: relevance and implications for public health.
Annu Rev Public Health. 2014;35:105–22.
57. Perera F, Herbstman J. Prenatal environmental exposures, epigenetics,
and disease. Reprod Toxicol. 2011;31(3):363–73.
58. Liu J, et al. Combined inhaled diesel exhaust particles and allergen
exposure alter methylation of T helper genes and IgE production in vivo.
Toxicol Sci. 2008;102(1):76–81.
59. Brand S, et al. DNA methylation of TH1/TH2 cytokine genes affects sen-
sitization and progress of experimental asthma. J Allergy Clin Immunol.
2012;29(6):1602–10.
60. Runyon RS, et al. Asthma discordance in twins is linked to epigenetic
modifications of T cells. PLoS ONE. 2012;7(11):e48796.
61. Kohli A, et al. Secondhand smoke in combination with ambient air
pollution exposure is associated with increasedx CpG methylation and
decreased expression of IFN-gamma in T effector cells and Foxp3 in T
regulatory cells in children. Clin Epigenetics. 2012;4(1):17.
Montrose et al. Allergy Asthma Clin Immunol (2017) 13:12
62. White GP, et al. CpG methylation patterns in the IFNgamma promoter in
naive T cells: variations during Th1 and Th2 differentiation and between
atopics and non-atopics. Pediatr Allergy Immunol. 2006;17(8):557–64.
63. Jones B, Chen J. Inhibition of IFN-gamma transcription by site-
specific methylation during T helper cell development. EMBO J.
2006;25(11):2443–52.
64. Lovinsky-Desir S, et al. DNA methylation of the allergy regulatory gene
interferon gamma varies by age, sex, and tissue type in asthmatics. Clin
Epigenetics. 2014;6(1):9.
Page 12 of 12
65. van den Brink GR, et al. Feed a cold, starve a fever? Clin Diagn Lab Immu-
nol. 2002;9(1):182–3.
66. Huang SL, Pan WH. Dietary fats and asthma in teenagers: analyses of the
first Nutrition and Health Survey in Taiwan (NAHSIT). Clin Exp Allergy.
2001;31(12):1875–80.
67. Heinrich J, et al. Allergic sensitization and diet: ecological analysis in
selected European cities. Eur Respir J. 2001;17(3):395–402.
Submit your next manuscript to BioMed Central
and we will help you at every step:
• We accept pre-submission inquiries
• Our selector tool helps you to find the most relevant journal
• We provide round the clock customer support
• Convenient online submission
• Thorough peer review
• Inclusion in PubMed and all major indexing services
• Maximum visibility for your research
Submit your manuscript at
www.biomedcentral.com/submit